Arm backend: Remove TOSA v0.80 from arm backend (pytorch#13040)

### Summary
This PR removes TOSA v0.80 from the arm backend. This includes updates
to:

- operator node visitors
- operator support 
- passes 
- test infrastructure
- tosa specification and utils
- removal of installation of v0.80 and patch scripts

Author: SaoirseARM

backends/arm/README.md

@@ -181,8 +181,8 @@ The Arm EthosU Backend should be considered a prototype quality at this point, l
 ## Current flows
 
 The EthosUBackend has a two stage process,
-- Compile to TOSA to rationalise the graph into known hardware support profiles. Currently this is to v0.80 TOSA BI with specific concern to a subset which gives support on Ethos-U55 and Ethos-U85, the target of the initial prototype efforts. This calls into the TOSABackend.
-- Lower via the ethos-u-vela compilation flow which takes TOSA v0.80 as an input and produces a low level commandstream for the hardware which is then passed via the delegate to the ethos-u-core-driver for direct execution.
+- Compile to TOSA to rationalise the graph into known hardware support profiles. Currently this is to v1.0 TOSA INT with specific concern to a subset which gives support on Ethos-U55 and Ethos-U85, the target of the initial prototype efforts. This calls into the TOSABackend.
+- Lower via the ethos-u-vela compilation flow which takes TOSA v1.0 as an input and produces a low level commandstream for the hardware which is then passed via the delegate to the ethos-u-core-driver for direct execution.
 
 The EthosUPartitioner is currenly used to ensure the operations converted are Ethos-U compatible, but will be extended to offer spec-correct TOSA Base inference and TOSA Main Inference generation in future.
 

backends/arm/_passes/arm_pass_manager.py

@@ -108,7 +108,7 @@ def _transform(self, graph_module: GraphModule):
         with TosaLoweringContext(self.tosa_spec):
             return self(graph_module).graph_module
 
-    def _tosa_080_BI_pipeline(self, exported_program: ExportedProgram) -> GraphModule:
+    def _tosa_INT_pipeline(self, exported_program: ExportedProgram) -> GraphModule:
         self.add_pass(FuseQuantizedActivationPass())
         self.add_pass(RemoveGetItemPass())
         self.add_pass(ConvertSplitToSlicePass())
@@ -164,7 +164,7 @@ def _tosa_080_BI_pipeline(self, exported_program: ExportedProgram) -> GraphModul
 
         return self._transform(exported_program.graph_module)
 
-    def _tosa_080_MI_pipeline(self, exported_program: ExportedProgram) -> GraphModule:
+    def _tosa_FP_pipeline(self, exported_program: ExportedProgram) -> GraphModule:
         self.add_pass(DecomposeMaskedFill())
         self.add_pass(DecomposeRoundPass())
         self.add_pass(DecomposeAcoshPass())
@@ -239,22 +239,12 @@ def _tosa_080_MI_pipeline(self, exported_program: ExportedProgram) -> GraphModul
 
         return self._transform(exported_program.graph_module)
 
-    def _tosa_1_0_int_quantized_pipeline(self, exported_program: ExportedProgram):
-        return self._tosa_080_BI_pipeline(exported_program)
-
-    def _tosa_1_0_fp_pipeline(self, exported_program: ExportedProgram):
-        return self._tosa_080_MI_pipeline(exported_program)
-
     def transform_to_backend_pipeline(self, exported_program: ExportedProgram):
         """Apply passes before transforming program to backend"""
-        if self.tosa_spec == TosaSpecification.create_from_string("TOSA-0.80.0+BI"):
-            return self._tosa_080_BI_pipeline(exported_program)
-        elif self.tosa_spec == TosaSpecification.create_from_string("TOSA-0.80.0+MI"):
-            return self._tosa_080_MI_pipeline(exported_program)
-        elif self.tosa_spec == TosaSpecification.create_from_string("TOSA-1.0+FP"):
-            return self._tosa_1_0_fp_pipeline(exported_program)
+        if self.tosa_spec == TosaSpecification.create_from_string("TOSA-1.0+FP"):
+            return self._tosa_FP_pipeline(exported_program)
         elif self.tosa_spec == TosaSpecification.create_from_string("TOSA-1.0+INT"):
-            return self._tosa_1_0_int_quantized_pipeline(exported_program)
+            return self._tosa_INT_pipeline(exported_program)
         else:
             raise NotImplementedError(
                 f"No pass pipeline implemented for {self.tosa_spec=}"

backends/arm/operator_support/convolution_support.py

@@ -21,8 +21,6 @@ class ConvolutionSupported(SupportedTOSAOperatorCheck):
     targets = [exir_ops.edge.aten.convolution.default]
 
     tosa_specs = [
-        TosaSpecification.create_from_string("TOSA-0.80+BI"),
-        TosaSpecification.create_from_string("TOSA-0.80+MI"),
         TosaSpecification.create_from_string("TOSA-1.0+INT"),
         TosaSpecification.create_from_string("TOSA-1.0+FP"),
     ]

backends/arm/operator_support/embedding_support.py

@@ -20,8 +20,6 @@ class EmbeddingSupported(SupportedTOSAOperatorCheck):
     targets = [exir_ops.edge.aten.embedding.default]
 
     tosa_specs = [
-        TosaSpecification.create_from_string("TOSA-0.80+BI"),
-        TosaSpecification.create_from_string("TOSA-0.80+MI"),
         TosaSpecification.create_from_string("TOSA-1.0+INT"),
         TosaSpecification.create_from_string("TOSA-1.0+FP"),
     ]

backends/arm/operator_support/index_select_support.py

@@ -18,8 +18,6 @@ class IndexSelectSupported(SupportedTOSAOperatorCheck):
     targets = [exir_ops.edge.aten.index_select.default]
 
     tosa_specs = [
-        TosaSpecification.create_from_string("TOSA-0.80+BI"),
-        TosaSpecification.create_from_string("TOSA-0.80+MI"),
         TosaSpecification.create_from_string("TOSA-1.0+INT"),
         TosaSpecification.create_from_string("TOSA-1.0+FP"),
     ]

backends/arm/operator_support/index_tensor_support.py

@@ -100,8 +100,6 @@ class IndexTensorSupported(SupportedTOSAOperatorCheck):
     targets = [exir_ops.edge.aten.index.Tensor]
 
     tosa_specs = [
-        TosaSpecification.create_from_string("TOSA-0.80+BI"),
-        TosaSpecification.create_from_string("TOSA-0.80+MI"),
         TosaSpecification.create_from_string("TOSA-1.0+INT"),
         TosaSpecification.create_from_string("TOSA-1.0+FP"),
     ]

backends/arm/operator_support/minmax_support.py

@@ -21,7 +21,6 @@ class MinMaxSupported(SupportedTOSAOperatorCheck):
 
     # TODO : "MLETORCH-718 : Quantization of indices in arm_quantizer"
     tosa_specs = [
-        TosaSpecification.create_from_string("TOSA-0.80+MI"),
         TosaSpecification.create_from_string("TOSA-1.0+FP"),
     ]
 

backends/arm/operator_support/pool_2d_support.py

@@ -43,8 +43,6 @@ class AvgPool2dSupported(SupportedTOSAOperatorCheck):
     ]
 
     tosa_specs = [
-        TosaSpecification.create_from_string("TOSA-0.80+BI"),
-        TosaSpecification.create_from_string("TOSA-0.80+MI"),
         TosaSpecification.create_from_string("TOSA-1.0+INT"),
         TosaSpecification.create_from_string("TOSA-1.0+FP"),
     ]
@@ -122,8 +120,6 @@ class MaxPool2dSupported(SupportedTOSAOperatorCheck):
     ]
 
     tosa_specs = [
-        TosaSpecification.create_from_string("TOSA-0.80+BI"),
-        TosaSpecification.create_from_string("TOSA-0.80+MI"),
         TosaSpecification.create_from_string("TOSA-1.0+INT"),
         TosaSpecification.create_from_string("TOSA-1.0+FP"),
     ]

backends/arm/operator_support/reduce_sum_support.py

@@ -19,8 +19,6 @@ class SumSupported(SupportedTOSAOperatorCheck):
     targets = [exir_ops.edge.aten.sum.dim_IntList]
 
     tosa_specs = [
-        TosaSpecification.create_from_string("TOSA-0.80+BI"),
-        TosaSpecification.create_from_string("TOSA-0.80+MI"),
         TosaSpecification.create_from_string("TOSA-1.0+INT"),
         TosaSpecification.create_from_string("TOSA-1.0+FP"),
     ]

backends/arm/operator_support/right_shift_support.py

@@ -27,8 +27,6 @@ class RightShiftSupported(SupportedTOSAOperatorCheck):
     ]
 
     tosa_specs = [
-        TosaSpecification.create_from_string("TOSA-0.80+BI"),
-        TosaSpecification.create_from_string("TOSA-0.80+MI"),
         TosaSpecification.create_from_string("TOSA-1.0+INT"),
         TosaSpecification.create_from_string("TOSA-1.0+FP"),
     ]